Ischemic stroke in young women

Initial and recurrent stroke.

Stroke was defined according to the criteria of the World Health Organization.18 The diagnosis of arterial ischemic stroke was confirmed in all patients by CT or MRI of the brain. The etiologic workup included MR angiography or ultrasonographic examination of cervical and cerebral arteries, 12-lead electrocardiography, transthoracic echocardiography, and routine laboratory tests. Transesophageal echocardiography and detailed coagulation studies, including antiphospholipid assessment, were performed in 60% of patients, and angiography in selected patients. Risk factors for stroke (hypertension, diabetes mellitus, current cigarette smoking, hypercholesterolemia, and current contraceptive use) were systematically recorded. The cause of stroke was classified as definite (large artery atherosclerosis, small artery disease, major cardiac source of embolism, other definite cause, and combination), uncertain, or undetermined, according to modified Trial of Org 10172 in Acute Stroke Treatment classification.19

The diagnosis of CVT was confirmed in all patients by MRI or angiography. Search for antinuclear antiphospholipid antibodies and detailed coagulation studies were performed in all patients.

Stroke recurrence was defined as a new neurologic deficit or an exacerbation of a previous deficit lasting more than 24 hours, according to Burn et al.20 The records of patients with a suspected stroke recurrence were reviewed by the coordinating center.

Statistical analysis.

Recurrence of stroke was assessed by life-table analysis according to Kaplan-Meier. Risk factors for recurrence were tested using proportional hazards models (Cox models). The hazard ratio and 95% CI were calculated for each factor; the hazard ratio may be interpreted as the relative risk of recurrence during the study period.

The person-time at risk (exposed person-time) during pregnancies and the 6 weeks after pregnancy was estimated from the numbers and average duration of spontaneous and induced abortions, stillbirths, and live births that have occurred in the population after the initial stroke. According to Kittner et al.,11 the duration of pregnancy was assumed to be 10 weeks for spontaneous and induced abortions, 28 weeks for stillbirths, and 38 weeks for live births. The postpregnancy period was defined as 6 weeks for each pregnancy outcome. Spontaneous abortion was defined as spontaneous fetal death before 20 weeks of gestation, and stillbirth was defined as spontaneous fetal death at 20 weeks of gestation or later. The person-time during which women in the population were not exposed (not pregnant or within 6 weeks after pregnancy) was estimated by subtracting the person-time calculated for exposure from the total person-time in the study population.

The recurrence rate (absolute risk and 95% CI) was estimated using the Poisson regression model. The relative risk of recurrence associated with pregnancy and postpartum exposure, adjusted for arterial or venous origin of initial stroke, was estimated using the Poisson regression model.

Study population.

A total of 489 patients were identified during the study period. Twenty patients were deceased at the time of the study (seven patients during the acute phase of their initial stroke and 13 before the beginning of the study [one vascular death and 12 nonvascular deaths]). None of these deaths were related to recurrent stroke or pregnancy.

Of the 469 surviving women, 9 declined to participate and 19 were lost to follow-up since the initial stroke. For five of them, we know only that they were still alive but were unable to obtain further information. A total of 411 women answered the questionnaire and the telephone interview. Information regarding stroke recurrence and subsequent pregnancies was obtained from the personal physician for 30 other women, leaving 441 women for the analysis. The baseline data, vascular risk factors, and causes of stroke for the 19 women who were lost to follow-up were similar to those of the 441 women included in the study.

The baseline characteristics and obstetric history of patients are given in table 1. A total of 260 women were monoparous (n = 74) or multiparous (n = 186) before stroke, with a parity ranging from one to eight. Seventeen patients had a definitive contraception before stroke (hysterectomy in three and tubal ligation in 14).

Initial ischemic stroke.

Of the 373 initial arterial ischemic strokes, 28 (7.5%) occurred during pregnancy or the puerperium (major cardiac source of embolism [n = 2], other definite cause [n = 11], and undetermined [n = 15]). Antithrombotic treatment was advocated for 330 patients (antiplatelet treatment [n = 216], warfarin [n = 110], and both [n = 4]). Sixteen women were discharged from the hospital without antithrombotic treatment, and information regarding prophylactic antithrombotic treatment was missing for 27 women. At the time of the interview (mean follow-up after initial ischemic stroke of 5 years), 232 (62%) women were taking antiplatelet drugs (n = 196), warfarin (n = 31), or both (n = 5).

Among the 68 initial CVT, nine occurred during the puerperium, two of which were associated with hematologic disease. Antithrombotic treatment was advocated for 54 patients (antiplatelet treatment [n = 8] and warfarin [n = 46]); information was missing for 14 patients. At the time of the interview, 23 (34%) women were taking antiplatelet drugs (n = 16), warfarin (n = 6), or both (n = 1).

Most patients (81.5%) had a minor or moderate residual handicap (Rankin score <3), including 42% of patients who had made a complete recovery from their stroke. Fifty-eight percent of women in our study returned to work. For 68 (15%) patients, a change in their marital status had occurred since the initial stroke (38 marriages or cohabitation, 25 divorces, and five deaths of the husband).

Recurrent strokes.

During a mean follow-up of 5 years (SD = 2.4; range 0.8 to 11.3), 13 women had a recurrent ischemic stroke. Two of these 13 women had pregnancy-associated strokes. All recurrent strokes were arterial in origin, and all were confirmed by neuroimaging studies. The initial stroke was arterial in 12 women and venous in another. No death or other vascular events occurred during the study period. The overall risk of recurrence was 1% (SD 0.5) within 1 year and 2.3% (SD 0.8) within 5 years (figure). Eleven factors were analyzed for an association with stroke recurrence, including age, vascular risk factors, subtype of initial stroke (arterial or venous), presence of a definite cause of stroke, and antithrombotic treatment advocated at time of discharge. The only statistically significant association was with an initial stroke of definite cause. The risk ratio was 8.6 for initial arterial stroke of definite cause (95% CI 1.9, 39.3; p = 0.005) and 7.9 for initial CVT of definite cause (95% CI 0.7, 88.7; p = 0.09).

• Download figure
• Open in new tab
• Download powerpoint

Figure. Kaplan-Meier curve shows the probability that patients will remain free of stroke recurrence.

Table 2 shows the absolute risk of recurrence outside and during pregnancy and the type of antithrombotic treatment at the time of recurrence. The relative risk of recurrence (see table 2), adjusted for the arterial or venous nature of initial stroke, was significantly higher for the postpartum period (risk ratio = 9.7; 95% CI 1.2, 78.9), but not during pregnancy itself (risk ratio = 2.23; 95% CI 0.3, 17.5).

Various regimens of antithrombotic drugs were advocated during subsequent pregnancies and the puerperium (table 3). Approximately half of women with a history of arterial ischemic stroke received antiplatelet treatment at least during a part of subsequent pregnancies or the puerperium, whereas most women who had CVT received no treatment during subsequent pregnancies or prophylactic heparinotherapy during the puerperium.

Reproductive history after stroke.

During a mean follow-up of 5 years (SD 2.4; range 0.8 to 11.3) after the initial stroke, 187 pregnancies occurred in 125 women, including 115 live births (61%), 37 induced abortions (20%), 30 miscarriages (16%), four ectopic pregnancies, and one stillbirth caused by retroplacental hemorrhage in a woman receiving low-dose aspirin (3%). The rates of subsequent pregnancies and of live births were 8.4 and 5 per 100 women-years. The mean delay between the initial stroke and the first subsequent pregnancy was 2.7 years (SD 1.84; range 1 to 7.9). Twenty-one pregnancies, including three live births, occurred in the first year after the initial stroke. Forty-six patients had two or more pregnancies after stroke. Of the 37 women whose initial stroke occurred during pregnancy or the puerperium (arterial ischemic stroke [n = 28] and CVT [n = 9]), 15 had one (n = 9), two (n = 5), or five (n = 1) subsequent pregnancies, including 11 live births. None of these pregnancies were associated with a recurrent stroke.

Among the 115 pregnancies that proceeded to delivery, 88 were vaginal deliveries (77%) and 19 (17%) were deliveries by cesarean section. The mode of delivery was unknown for eight patients. Peridural analgesia was used in 37 (32%) women.

Among the 352 women who answered the question regarding desire for pregnancy, 119 (34%) were unsatisfied with the number of pregnancies that they had after the initial stroke. Of these women, 91 (76.5%) had no subsequent pregnancies. The reasons given for not having the desired number pregnancies were related to the potential for stroke recurrence (n = 56, 47%), medical advice against pregnancy (n = 46, 39%), the existence of a residual deficit (n = 28, 24%), changes in sexual behavior (n = 12, 10%), hypofertility (n = 11, 9%) or modification of the family (n = 4, 3%). Women who had subsequent pregnancies were younger (mean age, 27.3 years versus 33.1 years; p = 0.0001) and less frequently hypertensive (4% versus 10%; p = 0.03) and more frequently had a Rankin score <3 (90% versus 78%, p = 0.015) than did women who had no subsequent pregnancy. The initial stroke was more frequently a CVT (23.2% versus 12.3%) than an arterial ischemic stroke (76.8% versus 87.7%; p = 0.004).

Few women (n = 39, 8.8%) used oral contraceptives after initial stroke (pure progestin pills [n = 27] or estroprogestative pills [n = 12]). Other methods of contraception, reported by 200 women, included intrauterine device (n = 119), condoms (n = 53), and vaginal foam (n = 28). Forty-four women had a definitive sterilization after stroke (tubal ligation [n = 37] or hysterectomy [n = 7]); in two other women, the husband had a vasectomy. One hundred thirty-five women used no contraceptive method (n = 135) after stroke; the method of birth control used was unknown for 21 women.

Recurrent strokes.

The overall risk of recurrent stroke in this study (1% within 1 year and 2.3% within 5 years) was comparable to that reported by others.21-23,25,26 The only prognostic factor significantly associated with recurrence was an initial stroke of definite cause, a finding consistent with the results of other series.23,28 Age, vascular risk factors, arterial or venous origin of stroke, and antithrombotic treatment advocated at time of discharge did not influence the risk of recurrence, but these observations are limited by the small number of events.

Only two of the 13 recurrent strokes observed in our study occurred during subsequent pregnancies or the puerperium, and these two strokes were related to rare definite causes of stroke (essential thrombocytemia and primary antiphospholipid syndrome). It is noteworthy that no woman whose initial stroke occurred during pregnancy had a recurrent stroke during subsequent pregnancies. The postpartum period, not the pregnancy itself, was associated with an increased relative risk of stroke recurrence. These results are in agreement with our previous collaborative study, based on surveillance at 63 public maternity departments in Ile de France, showing that the risk of ischemic stroke per day is higher during the postpartum period than during any trimester of pregnancy.10 Likewise, in the study by Kittner et al.11 dealing with all female patients aged 15 to 44 years discharged from any of 46 hospitals in central Maryland and Washington DC in 1988 or 1991, the relative risk of ischemic stroke was increased in the 6 weeks after delivery (relative risk = 8.7; 95% CI 4.6, 16.7), but not during pregnancy itself (relative risk = 0.7; 95% CI 0.3, 1.6). The higher relative risk of first or recurrent ischemic stroke associated with the postpartum period suggests a causal role for the large decrease in blood volume or the rapid changes in hormonal status that follow a live birth or stillbirth, perhaps by means of hemodynamic, coagulative, or vessel wall changes.11,15

There is no consensus on the risk–benefit ratio of antithrombotic treatment to prevent recurrent ischemic stroke during pregnancy.16,29 The various antithrombotic drug regimens that were used during pregnancy and the puerperium in our study (see table 3) reflect this lack of guidelines. If prophylactic antithrombotic treatment proves to be beneficial, it is likely to be so during late pregnancy or the puerperium.

Reproductive history.

A total of 187 pregnancies occurred during the 5-year follow-up after initial stroke. The annual incidence of pregnancy and the distribution of the various outcomes of pregnancy observed in our study (see results) are similar to those reported in women of reproductive age in France30 (annual incidence of pregnancy, nine per 100 women; live births, 65%; induced abortions, 18%; miscarriages, 15%; other, 2%). These results cannot be directly compared because many factors, such as parity, socioeconomic status, and fertility, can influence these figures.31

Our study focused on the impact of stroke on the desire for pregnancy. An interesting finding is that more than one-third of women stated that they were unsatisfied with the number of pregnancies that they had after stroke. Desire for pregnancy, however, is a complex notion for which there is no standardized and validated questionnaire. It depends on multiple factors, including age, number of children, general health of the woman, and neurologic impairment after stroke. In our study, women who had subsequent pregnancies were younger and had a less severe deficit than did those who had no subsequent pregnancies. The main reasons that women gave for not having subsequent pregnancies after stroke were, in order of decreasing frequency, concern of recurrent stroke during pregnancy, medical advice against pregnancy, and residual deficit. The belief that subsequent pregnancy was associated with a high risk of recurrent stroke was sometimes carried out by general physicians or obstetricians who warned women to avoid pregnancy. Some women were even encouraged by their physician to undergo sterilization or induced abortion solely because they had suffered a stroke and such advice was not related to the presence of a definite cause of stroke. The low risk of recurrent stroke during pregnancy in our study does not support these recommendations and suggests that the occurrence of an ischemic stroke should not contraindicate a subsequent pregnancy. The final decision depends on several factors, including the cause of initial stroke, the desire for pregnancy, and the residual deficit. Although neurologic impairment was more frequent in women who had no pregnancy after stroke, only 46% of women who stated that residual deficit was the reason for not considering pregnancy had a Rankin score >2, which corresponds to a moderate or severe handicap. This finding is probably explained by physical or emotional limitations that could not be measured by the Rankin score. Hypofertility or sexual behavior changes since stroke were rarely mentioned.

No data and no guidelines are available for the management of labor and delivery in women with a history of ischemic stroke. In our study, cesarean section delivery and fetal outcomes were similar to those reported in the French population,32 although direct comparison is difficult. Peridural analgesia has been used in a small proportion of vaginal deliveries without any adverse event. It has not been performed in many cases because of perceived complications rather than documented risks associated with this technique in women who had previous ischemic stroke. In our experience, a history of ischemic stroke is not sufficient in itself to contraindicate peridural analgesia or to recommend cesarean section delivery.

Concerning contraception after stroke, our study shows that most women used, as expected, mechanical methods of birth control. Only a few used oral contraceptives. No recurrent stroke occurred in these women, but the small number of events precludes any conclusion on the risk of recurrent stroke associated with oral contraception.

Study limitations.

Our study has several potential limitations. First, the results are based on a small number of observed events that resulted in large CIs for the relative and absolute risk of recurrent stroke. Because of the small number of events, we cannot exclude the possibility that some nonsignificant results could become significant with increases in sample size. Second, recurrent cerebrovascular events and subsequent pregnancies were not prospectively recorded. As a consequence, we may have missed some events that occurred during follow-up. Most women in our study, however, were regularly followed up by their neurologist, and all medical records were reviewed. In addition, strokes in young women are always unusual events that engender a high degree of concern and lead to hospitalization for nearly all patients. Therefore, it is unlikely that significant cerebrovascular events were missed. It also seems unlikely that subsequent pregnancies, in particular those that resulted in a live birth, may have been not mentioned by the women. Because some cases of spontaneous or induced abortions may have been unrecognized or unreported, it is possible that we may have underestimated the risk of recurrent stroke associated with abortion. Third, 19 (4%) women were lost to follow-up. A selection bias could have occurred if these women were more likely to have recurrent strokes associated with pregnancy. The baseline data, vascular risk factors, and causes of stroke, however, were similar to those of women included in the study. Because their general physicians had no follow-up information since the initial stroke, it is probable that these patients moved away from the area.